Qingyun Research Intern · Top Talent Internship

Large-scale video foundation model pre-training and post-training, SAGE-GRPO, agentic data recipes, emotion data, agentic RL, and reward systems for image and video generation.

Researcher & Algorithm Engineer

Developed autoencoders and experimental pipelines for image and video generation, and studied frontier video-model architectures and training paradigms.

Computer Vision Algorithm Intern

Developed multilingual forced alignment across 26 languages for HeyGen’s first-generation avatar pipeline and researched audio- and video-driven talking-head generation.

Machine Learning Algorithm Intern

Built a millimeter-wave-radar sleep-status prediction system with LightGBM and improved accuracy from 72% to 86%.

PhD Student & Researcher · Generative Models, Multimodality, Movie Generation

Research on agentic and multimodal visual generation, foundation-model post-training, and machine-assisted movie creation.

Research Assistant · Generative Models & Multimodality

Worked on LLM-assisted visual generation, FineCLIPER, and high-fidelity text-guided 3D face generation with diffusion models and neural radiance fields.

Research Assistant · Generative Models & Text-to-3D

Studied text-to-3D architectures and acceleration strategies, combining efficient generation with statistical machine learning.

Research Assistant · Generative Models & Super Resolution

Developed a latent-diffusion and VQGAN system for super resolution and explored modified latent diffusion for visual classification.

Undergraduate Researcher · HCI, Computer Vision & EEG

Built EEG- and gaze-driven interaction prototypes and led a deep-learning project on cognitive-perception fatigue-driving detection.